Purpose As reliability of raster stereography was proved only for sagittal plane parameters with repeated measures on the same day, the present study was aiming at investigating variability and reliability of back shape reconstruction for all dimensions (sagittal, frontal, transversal) and for different intervals. Methods For a sample of 20 healthy volunteers, intraindividual variability (SEM and CV %) and reliability (ICC ± 95 % CI) were proved for sagittal (thoracic kyphosis, lumbar lordosis, pelvis tilt angle, and trunk inclination), frontal (pelvis torsion, pelvis and trunk imbalance, vertebral side deviation, and scoliosis angle), transversal (vertebral rotation), and functional (hyperextension) spine shape reconstruction parameters for different test-retest intervals (on the same day, between-day, between-week) by means of video raster stereography. Results Reliability was high for the sagittal plane (pelvis tilt, kyphosis and lordosis angle, and trunk inclination: ICC [ 0.90), and good to high for lumbar mobility (0.86 \ ICC \ 0.97). Apart from sagittal plane spinal alignment, there was a lack of certainty for a high reproducibility indicated by wider ICC confidence intervals. So, reliability was fair to high for vertebral side deviation and the scoliosis angle (0.71 \ ICC \ 0.95), and poor to good for vertebral rotation values as well as for frontal plane upper body and pelvis position parameters (0.65 \ ICC \ 0.92). Coefficients for the between-day and between-week interval were a little lower than for repeated measures on the same day. Variability (SEM) was less than 1.5°or 1.5 mm, except for trunk inclination. Relative variability (CV) was greater in global trunk position and pelvis parameters (35-98 %) than in scoliosis (14-20 %) or sagittal sway parameters (4-8 %). Conclusions Although we found a lower reproducibility for the frontal plane, raster stereography is considered to be a reliable method for the non-invasive, three-dimensional assessment of spinal alignment in normal non-scoliotic individuals in the sagittal plane and partly for scoliosis parameters, which fulfils scientific as well as practical recommendations for spine shape screening and monitoring, but cross-sectional or follow-up effect analyses should take into account the degree of reliability differing in various spine shape parameters. Further investigations should be conducted to analyse reliability in scoliosis patients with differing spinal deformities.